\documentclass[a4paper, 11pt]{article}

\usepackage{amsfonts}
\usepackage{amsmath}
\usepackage{amsthm}
\usepackage{appendix}
\usepackage{bm}
\usepackage{booktabs}
\usepackage[usenames, dvipsnames]{color}
\usepackage{graphicx}
\usepackage{epstopdf}
\epstopdfsetup{update}
\usepackage{helvet}
\usepackage{hyperref}
\usepackage{indentfirst}
\usepackage{lscape}
\usepackage{longtable}
\usepackage{pdflscape}
\usepackage{morefloats}
\usepackage{natbib}
\bibliographystyle{aea}
\usepackage{setspace}
\usepackage{subcaption}
\usepackage[capposition=top]{floatrow}
\usepackage{subfloat}
\usepackage[latin1]{inputenc}
\usepackage{tikz}
\usepackage{eurosym}
\usepackage[margin=0.86in]{geometry}


 \setlength{\parskip}{1em}

\hypersetup{
    colorlinks=true,
    linkcolor=black,
    citecolor=black,
    filecolor=black,
    urlcolor=black
}

\title{\textbf{Online Appendix for ``On the Value of Birth Weight''}\thanks{\scriptsize{This experiment documented in this paper has passed ethical approval at the Oxford Centre of Experimental Social Sciences (CESS), and been registered as project ETH-160128161. We thank the editor James Fenske, two anonymous referees at the Oxford Bulletin of Economics and Statistics, \'{A}ureo de Paula, two anonymous referees at Review of Economic Studies, and participants in seminars at the University of Exeter and the University of Surrey, and at the Royal Economic Society (RES) Annual Conference 2019 (University of Warwick) and European Society for Population Economics (ESPE) Annual Conference 2019 (University of Bath) for helpful comments and suggestions. eplication materials are available at the Harvard Dataverse, DOI: \url{https://doi.org/10.7910/DVN/IWINJN}. Any errors are our own.}}}

\author{\small{Damian Clarke} \\ \small{Universidad de Chile \& IZA} \and \small{Sonia Oreffice} \\ \small{University of Exeter, HCEO \& IZA}  \and \small{Climent Quintana-Domeque} \\ \small{University of Exeter, GLO, HCEO \& IZA}}

\date{\today}
\begin{document}
\begin{spacing}{1.4}
\maketitle

\begin{abstract}
This document contains the online appendix for ``On the Value of Birth Weight''.
\end{abstract}

 \end{spacing}

\newpage

\clearpage


 
\setcounter{table}{0}
\renewcommand{\thetable}{A\arabic{table}}
\setcounter{figure}{0}
\renewcommand{\thefigure}{A\arabic{figure}}

\appendix
\section*{Additional Figures and Tables}

\begin{figure}[htpb!]
  \begin{center}
    \caption{Geographical Coverage of Respondents}
    \label{geography}
    \includegraphics[scale=0.6]{./results/Summary/surveyCoverage-orig.eps}
  \end{center}
  \floatfoot{\textsc{Notes:} The survey sample consists of 2,005 respondents.  The final estimation sample consists of 1,894 respondents after removing respondents whose geographic IP suggested a non-US location (72 respondents, 3.6\%), those who failed to respond that their educational attainment was identical at the beginning and end of the survey 26 respondents, 1.3\%, and those who completed the discrete choice experiment in under two minutes (16 respondents, 0.8\%).}
\end{figure}


%A1
\begingroup
\setlength{\LTleft}{-20cm plus -1fill}
\setlength{\LTright}{\LTleft}
\begin{longtable}{lccc}
  \caption{Geographical Coverage of Respondents} \label{tab:cover} \\
  \hline
  State Name & Percent & Percent & Difference \\
  & MTurk      & Census Bureau &           \\ \hline \endhead
  \input{./results/Summary/GeographicCoverage.tex}
  \hline
  \multicolumn{4}{p{10.4cm}}{{\footnotesize\textsc{Notes:} Columns present the percentage of respondents from the MTurk sample, the percentage of residents according to US Census Bureau records (2015), and the difference between the percentage of MTurk respondents and residents.}}
\end{longtable}
\endgroup


%A2
\begin{table}[htpb!]
  \caption{Comparison of MTurk Respondents with American Community Survey}
  \label{tab:compACS}
  \begin{tabular}{lcccccc} \toprule
    Variable & \multicolumn{2}{c}{MTurk Sample} &\multicolumn{2}{c}{ACS Sample} & \multicolumn{2}{c}{Difference} \\ \cmidrule(r){2-3}\cmidrule(r){4-5}\cmidrule(r){6-7}
    & Mean & Std.\ Dev.\ & Mean & Std.\ Dev.\ & Diff. & t-Stat. \\ \midrule
    \input{./results/Summary/ACScomp.tex}
    \bottomrule
    \multicolumn{7}{p{14.0cm}}{{\footnotesize Notes: MTurk sample refers to the
        full estimation sample from the Discrete Choice Experiments implemented
        on Amazon Mechanical Turk and ACS Sample refers to the 2015 American Community
        Survey universe of respondents between 18 and 75 years.  Sample weights are
        used when calculating ACS summary statistics.  The absolute difference between
        means is displayed in the column labelled Diff, and a t-statistic associated
        with a t-test of equality of means is displayed in the t-Stat.\ column.}}
  \end{tabular}
\end{table}


\begin{figure}[htpb!]
  \begin{center}
    \caption{Total Family Income: MTurk Sample and ACS Sample (2015)}
    \label{incomeDesc}
    \includegraphics[scale=0.9]{./results/Summary/income.eps}
  \end{center}
  \floatfoot{\textsc{Notes}: Income levels are reported in bins.  In the MTurk
    survey, respondents reported their binned family income, while in ACS, absolute
    family income is reported. The full analysis sample is compared with the full ACS
    sample aged between 18-75.}
\end{figure}



\clearpage



\begin{figure}[htpb!]
  \begin{center}
    \caption{Discrete Choice Experiment Framing}
    \label{DCE-frame}
    \includegraphics[scale=0.7]{./results/auxiliary/Q1.png}
  \end{center}
 \end{figure}

\begin{figure}[htpb!]
  \begin{center}
    \caption{Discrete Choice Experiment Options}
    a    \label{DCE-options}
    \includegraphics[scale=0.7]{./results/auxiliary/Q2.png}
  \end{center}
 \end{figure}


\begin{figure}[htpb!]
  \begin{center}
    \caption{Discrete Choice Experiment Example}
    \label{DCE-example}
    \includegraphics[scale=0.7]{./results/auxiliary/Q3.png}
  \end{center}
 \end{figure}

\clearpage
\begin{figure}[htpb!]
  \begin{center}
    \caption{Birth Weight from Administrative Data}
    \label{bwt-nvss}
    \includegraphics[scale=0.9]{./results/auxiliary/birthweight.eps}
  \end{center}
  \floatfoot{\textsc{Notes:} Full birth weight distribution from all US births occurring in 2013 observed from NVSS birth certificate data (values below 500 grams or above 5000 grams are removed for display purposes). 84.09\% of all births fall in the ``normal'' birth range of 2500 to 4000g.  Of non-normal birth weights, 8.02\% are low birth weight ($<$ 2,500 grams), and the remaining 7.89\% were large ($>$ 4,000 grams).}
\end{figure}


\begin{figure}[htpb!]
  \begin{center}
    \caption{Discrete Choice Experimental Results with Categorical Costs}
    \label{DCE-full-samp}
    \includegraphics[scale=0.9]{./results/Figures/Conjoint_Sample.eps}
  \end{center}
  \floatfoot{\textsc{Notes:} Refer to note to Figure 2.  This figure is based
    on an identical sample, however now using a categorical, rather than a linear, measure of
    costs.}
\end{figure}


\begin{figure}[htpb!]
  \begin{center}
    \caption{Discrete Choice Experimental Results (Full Sample)}
    \label{DCE-sampFull}
    \includegraphics[scale=0.9]{./results/Figures/Conjoint_All_continuous.eps}
  \end{center}
  \floatfoot{\textsc{Notes:} Refer to notes to Figure 2.  This figure is
    identical, however now also including the $\sim 5\%$ of the sample removed for
    failing consistency checks.}
\end{figure}


\input{./results/Regressions/conjointWTP-weighted.tex}




\begin{landscape}
  \input{./results/Regressions/conjointWTP-interactions.tex}
\end{landscape}

\begin{table}[htpb!]
  \begin{center}
    \caption{Summary Statistics of Respondents by Parental Status}
    \label{tab:sumstatsPNP}
    \begin{tabular}{lccccc} \toprule
      &           N&        Mean&  Std.\ Dev.&         Min&         Max\\ \midrule
      \multicolumn{6}{l}{\textbf{Panel A: Parents}} \\
      \input{./results/Summary/MTurkSum-parent-clean.tex}
      \multicolumn{6}{l}{\textbf{Panel B: Non-Parents}} \\
      \input{./results/Summary/MTurkSum-nonparent-clean.tex}
      \bottomrule
      \multicolumn{6}{p{12.2cm}}{{\footnotesize\textsc{Notes:} Full summary
          statistics and notes are provided in Table 1.
          Here identical values are provided for parents (Panel A) and non-parents
      (Panel B) separately.}}
    \end{tabular}
  \end{center}
\end{table}




\clearpage
\begin{landscape}
  \input{./results/Regressions/conjointGroups_het.tex}
\end{landscape}


\clearpage
\input{./results/Regressions/conjointWTP-A.tex}


\clearpage
\begin{landscape}
\input{./results/Regressions/conjointGroupsDemog.tex}
\end{landscape}

\clearpage




\begin{landscape}
  \input{./results/Regressions/WTP-mixedlogit-clean.tex}
\end{landscape}

\begin{figure}[htpb!]
  \begin{center}
    \caption{Distribution of Willingness to Pay in the Population}
    \label{mixWTP-dist}
    \includegraphics[scale=0.77]{./results/Figures/WTPdist-all.eps}
  \end{center}
  \floatfoot{\textsc{Notes:} Distribution of WTP among all respondents is estimated
    from a mixed logit model using a linear specification for birth weight, and the
    conditioning of individual taste (COIT) procedure described in
    \citet{ReveltTrain2000}.  All respondents are used.}
\end{figure}

\begin{figure}[htpb!]
  \begin{center}
    \caption{Willingness to Pay and Proportion Positively Valuing Birth Weight (Mixed Logit)}
    \label{mixWTP-NP}
    \includegraphics[scale=0.77]{./results/Figures/WTP_mixed_all.eps}
  \end{center}
  \floatfoot{\textsc{Notes:} A mixed logit specification is estimated, however now birth weight is allowed to enter non-parametrically.  Willingness to Pay for each component is with respect to the baseline birth weight of 2,500 grams.  The Percent Preferring Birth weight refer to the percentage of all respondents who positively value a given weight versus the baseline category.  Refer to notes to Table \ref{WTPmix} for additional details.}
\end{figure}

\end{document}
